Ryzen was being reported at higher Ghz than it really ran with HPET disabled. I believe some benchmark results like Handbrake would report incorrectly at well. Though they were patched so you could get correct results with HPET disabled about a month after Ryzen 1 release, IIRC.
So this is very similar to those early Ryzen bugs. But worse, and possibly intentional.
Intel may be intentionally obfuscating how much Spectre/Meltdown affected them with this loophole, and telling reviewers to bench with HPET disabled.
Assuming it's true, that is, but I think it may very well be. Testing to verify is very difficult since you need to measure the frames of a synced monitor compared to what the FPS counter claims.